This paper is concerned about testing whether a cross-covariance matrix deviates\nfrom a pre-assigned one or not. For this purpose, a new test statistic is\nconstructed based on the Frobenius norm of the difference between the\nsample cross-covariance matrix and the pre-assigned matrix. The test is implemented\nby applying the parametric bootstrap scheme. We conduct a simulation\nstudy to examine the performance of the test and compare it with\nother competitive tests. As multiple simulation examples show, our empirical\npowers are clearly superior to others in detecting any deviation of the\ncross-covariance from the pre-assigned matrix. In addition, the proposed\ntest is insensitive to non-cross-covariance elements in the covariance matrix.\nAs an illustration, we also investigate its performance in testing pairwise\ntime-reversibility.
Loading....